Multi-view video coding (MVC) is the compression framework for the encoding of multi-view sequences. A multi-view video coding sequence is a set of two or more video sequences that capture the same scene from a different view point.
Motion skip mode has been proposed to improve the coding efficiency for multi-view video coding. A basis of this proposal is the assumption that there is a similarity with respect to the motion between the neighboring two views, therefore coding efficiency can be achieved by exploiting this redundancy. Motion skip mode infers the motion information, such as macroblock type, motion vector, and reference indices, directly from the corresponding macroblock in the neighboring view at the same temporal instant. The method is decomposed into two stages, namely a search for the corresponding macroblock and a derivation of motion information. In the first stage, a global disparity vector (GDV) is used to indicate the corresponding position (macroblock) in the picture of a neighboring view. The global disparity vector is measured by the macroblock-size of units between the current picture and the picture of the neighboring view. The global disparity vector can be estimated and decoded periodically such as, for example, every anchor picture. In that case, the global disparity vector of a non-anchor picture is interpolated using the recent global disparity vectors from the anchor picture. In the second stage, the motion information is derived from the corresponding macroblock in the picture of the neighboring view, and the motion information is copied in order to be applied to the current macroblock. Motion skip mode is disabled in the case when the current macroblock is in a picture of the base view or is in an anchor picture, because the proposed method exploits the picture from a neighboring view to present another way for the inter prediction process.
To signal the use of motion skip mode at the encoder to the decoder, a flag referred to as motion_skip_flag is included in the head of macroblock layer syntax for multi-view video coding. If motion_skip_flag is turned on, then the current macroblock derives macroblock type, motion vector, and reference indices from the corresponding macroblock in the neighboring view. Motion skip mode needs to know which reference is the inter-view reference for the current picture in order to infer the motion information, such as macroblock type, motion vector, and reference indices, directly from the corresponding macroblock in the neighboring view at the same temporal instant, but does not predict any pixel value from the inter-view reference picture.